Match ups with McCain I understand. But this? Somebody explain, please.

Sometimes I suspect that some small percentage of people in these polls actually don't recognize his name (there's always some of them) and, when polled on whether he's a Muslim, just guess based on the name. (And some others actually believe it, of course.)

Meaningless.

This is the year of the Blue Tidal Wave. The "unlikely" voters. Many don't have land lines and don't answer surveys, FWIW.

Maybe something to do with the phrasing of the question? I can't find the actual wording used. Maybe one asked "who do you like better" and the other "who do you want to be the nominee" or something like that.

That's typical.

Now add in factors for the non-randomness - only land lines phoned, only people with phones called, not everyone votes the way they say ... and you get an even larger number of incorrect polls, @6.

Then add in all the "unlikely" voters this year and you realize all your polling methods are pretty useless - normally we count 3s and 4s, not 1s and 2s ... but this year most voters are people who rarely if ever vote ...

The Newsweek question was:

Who would you MOST like to see nominated as the Democratic Party’s presidential candidate this year… Hillary Clinton (or) Barack Obama? (Choices Rotated)

I can't find the Gallup wording. Anyone?

Margin of error is based on 95% confidence, but that doesn't mean everything outside the margin of error is equally likely. The polls are just too different. You could base this on three standard deviations instead of two, which is 99.3% confidence, and the margin of error still would not account for the difference in the two polls.

The other 10 percent use wikipedia.